Outside the Box:

Unconventional Monetary Policy in

the Great Recession and Beyond
Ken Kuttner
Williams College

Hutchins Center on Fiscal and Monetary Policy at Brookings

October 17, 2018
 Prong #1: Quantitative Easing (QE)
4500
MBS • Four programs: the three
4000 5+ years
Large Scale Asset Purchases
1-5 years LSAP3
3500
Bills (LSAPs) plus the Maturity
3000 MEP
Extension Program (MEP).
2500 LSAP2 Collectively, “QE.”
billion

LSAP1
2000
• Different asset mix, effects
1500
Crisis on the balance sheet.
1000

500
• Far larger in size than with
0
conventional policy.
2007 2008 2009 2010 2011 2012 2013 2014 2015
Prong #2: Forward Guidance (FG)
 December 18, 2008: “the Committee anticipates that weak economic
conditions are likely to warrant exceptionally low levels of the federal
funds rate for some time.”
 March 18, 2009: “…exceptionally low levels of the federal funds rate for
an extended period.”
 September 21, 2011: “…exceptionally low levels for the federal funds rate
at least through mid-2013.”
 December 12, 2012: “…this exceptionally low range for the federal funds
rate will be appropriate at least as long as the unemployment rate
remains above 6-1/2 percent…”
What the Fed’s UMP was not
 Emergency asset asset purchase and lending programs
(Commercial Paper Funding Facility, etc.) “Lender of last
resort” actions, not monetary policy.
 Bank of Japan style quantitative easing, which targeted the
volume of bank reserves (“current account balances”), rather
than the asset side of the balance sheet.
Why is it so hard to measure UMP’s impact?
 What would have happened in the absence of the policies?
 A skeptic would say:
• “Interest rates fell after QE and FG… but maybe they would have
fallen anyway, given the weakening economy.”
• “The economy continued to contract, despite QE and FG.
Therefore, the policies were ineffective.”
• “The economy eventually recovered, but maybe it would have
done so on its own.”
Additional complications
 The short track record—seven years, not even a full business
cycle.
 Poorly understood transmission mechanism.
 The policies’ heterogeneity—differences in the kinds of
assets purchased, nature of announcements.
 Lots of other things (e.g. the financial market’s near death
experience) were happening at the same time.
Types of evidence on UMP’s effectiveness

 Impact on interest rates

• Event studies
• Econometric term structure models
 Effects on the real economy
• Aggregate macroeconomic models
• Micro-level studies
 What do the event studies tell us?
20
The bars represent the
10
changes in the 10-year
0
Treasury yield on
Basis points

-10
announcement days.
-20

-30

-40
LSAPs
-50
3/18/2009
-60
LSAP1 LSAP2 MEP LSAP3
100 bp 18 bp 27 bp 11 bp

 The effects of LSAP 1 are huge…

 …but quite modest for subsequent asset purchase programs.
 Can we disentangle QE from FG?
20
The pink bars are the
10
LSAPs, the blue bars
0
are the FG
Basis points

-10
announcements.
-20

-30

-40
LSAPs
Forward guidance
-50
3/18/2009
-60
LSAP1 LSAP2 MEP LSAP3
100 bp 18 bp 27 bp 11 bp

 The days with the largest LSAP1 reactions were also days with FG
announcements.
Other event study caveats
 Were the effects of the early LSAP1 announcements due to a
“market functioning” effect?
 Maybe the announcements were anticipated?
 How persistent were the effects?
Econometric term structure modeling
 A economically rigorous (but opaque) way to estimate the
effects of asset quantities on bond yields.
 Unlike event study method, it does not depend on the
markets being “surprised.”
 Ballpark estimates: 40-50 bp for the three LSAPs, 20 bp for
the MEP.
 This is largely through a reduction in the “term premium”
(but perhaps expectations to some extent).
 Estimates of the 10-year term premium
6
QE1 yield Black = 10-year
5 premium Treasury yield, blue =
QE2
MEP term premium.
4
QE3
3
Difference between
percent

2 the two is the average

1
expected level of
short-term rates.
0

-1
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

 The reduction in the supply of 10-year bonds led to a record low

Treasury yield (1.5% in July 2012) and a negative term premium.
Did UMP have real effects?
 Macro method: feed bond yield reductions into
macroeconomic models. Typical estimate: a 1 percentage
point reduction in the unemployment rate.
 Micro method: look for different effects across firms or banks.
• Firms with more reliance on long-term debt benefited more from
the MEP relative to those using more short-term debt.
• Banks with more exposure to MBS benefitted more from LSAP1 and
LSAP3 relative to those with less exposure.
Unintended consequences?
 Inflation?
 Spillovers—especially to emerging markets.
• Not specific to unconventional policy…
• …and surely preferable to a more protracted US recession.
 Excessive risk taking?
• What is “excessive”?
• “Micro” risk taking could reduce “macro” risk.
Should the Fed use UMP “next time”?
 Did it work?
• Each study has its limitations and the magnitudes are uncertain…
• …but by the “preponderance of evidence” standard, yes.
 Did it have any downsides?
• Nothing major…
• …although there is some concern about “too low for too long.”
 Why not?
What could/should UMP look like next time?
 UMP should be predictable and rule-like, like LSAP3. FG could
be a complement.
 Separate management of short- and long-term interest rates
creates an additional policy tool—what to do with it?
 Could FG be used independently of QE? Maybe—but it’s
tricky to send the right message.